I'm having two significant problems with my automatic merge process (described below). Can anyone give me advice on how to improve the process or recommend other well known practices? The process to merge branch_A into branch_B: 1) check-out branch_A 2) find the checkpoint tag from the last branch_A to branch_B merge and assign to old_checkpoint 3) increment the numeric portion of old_checkpoint and assign to new_checkpoint 4) tag branch_A with new_checkpoint 5) check-out branch_B 6) find any files/directories that exist in branch_A, but not in branch_B and add them to branch_B. 7) In the branch_B directory do: cvs update -d -j old_checkpoint -j branch_B 8) If there were conflicts remove new_checkpoint and exit with error 9) otherwise commit changes in the branch_B directory The problems I'm having are: 1) Some changes in branch_A are not showing up in branch_B. I suspect this is because the old_checkpoint tag was somehow left behind from a failed merge. 2) Step 6 of the merge process is adding files that were intentionally removed from branch_B, but not from branch_A. I shouldn't have to perform step 6 though. CVS should automatically pick up new files and directories as well as removing old ones. However this does not seem to happen. Any help is appreciated. --Craig Green _______________________________________________ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
